Task-model choice
Choose supported cloud providers, compatible endpoints or a separately running local model server for the high-level agent.
THE FIELD GUIDE / 10 — Skills, models & connections
Start with Jarvis. Make room for your way.
Choose the task model, connect MCP tools and teach reusable ways of working.
Create a reusable review skill that checks public contracts, small-screen layouts and relevant tests.
JARVIS / IN EVERYDAY LIFEIllustrated possibilities. Your tools do the work.
01 / THE POSSIBILITIESA little more
A little more
within reach.
Your assistant should fit the tools and habits you already have. Jarvis exposes provider configuration, MCP connections, Python extensions, reusable skills and agent profiles. The useful combination is the one you actually configure and can inspect.
See it in your dayMCP tool connections
Configure servers, discover the tools they expose and invoke them with the credentials and access each service requires.
Skills & Python extensions
Add reusable instructions, custom tools, lifecycle hooks, prompt templates and specialist agent definitions.
Tool permissions & profiles
Configure tool access and path rules for the agent. Approval behavior depends on the active runtime and permission configuration.
Authenticated local services
The voice and task brains communicate through A2A; desktop tools use their local service. These retained services support integrations, not a bundled chat website.
02 / PUT IT INTO YOUR DAYOrdinary requests.
Ordinary requests.
Useful little transformations.
Follow the thought from the first sentence to the result. Make the names, files, and timing your own.
THE EVERYDAY / WORKFLOW 01
Turn your review habits into a skill
You keep explaining the same review priorities at the start of each project.
START WITH THIS
“Create a reusable review skill that checks public contracts, small-screen layouts and relevant tests.”
- 01
Start with the context
Define the scope, expected checks and useful output format in a reusable skill document.
- 02
Let the work unfold
Make the skill available through Jarvis's skill system, then ask the task agent to use it on a specific project.
- 03
Check the result
Review the first result and refine the instructions. The skill guides tool use; it does not install the project's runtime or grant access.
THE RESULT TO LOOK FOR
A reusable working method that can be applied when relevant. Quality still depends on the instructions, model and actual tool evidence.
THE EVERYDAY / WORKFLOW 02
Bring a configured service into the workflow
A service you use exposes an MCP interface and you want Jarvis to use it deliberately.
START WITH THIS
“Show the tools available from my configured MCP server.”
- 01
Start with the context
Configure the server's command or transport and the authentication it requires.
- 02
Let the work unfold
Jarvis connects and discovers the actual tools available from that server.
- 03
Check the result
Choose a specific operation and inspect its returned result. Availability follows that connection's permissions and service capabilities.
THE RESULT TO LOOK FOR
A discovered and usable tool connection when configured successfully. An installed skill mentioning a service is not proof of an authenticated connection.
THE EVERYDAY / WORKFLOW 03
Use the task model that fits your setup
You want the deeper reasoning work to use your own running inference endpoint.
START WITH THIS
“Configure my local model server for the high-level task agent.”
- 01
Start with the context
Install and start the model server separately, then select the corresponding supported provider configuration.
- 02
Let the work unfold
Set the endpoint, model and any required credentials through Jarvis's model setup.
- 03
Check the result
Try a bounded task with the tools you need. Keep Gemini Live configured for the voice side of the two-brain runtime.
THE RESULT TO LOOK FOR
A configured task-model connection if the endpoint and model are compatible. Tool quality, context limits and image support vary by provider.
Choose a provider
Use the model configuration command and the matching credentials. Local servers need their own installed model and running process.
uv run jarvis modelConfigure integrations
Use Jarvis setup and the source documentation for each MCP server, extension or gateway adapter.
uv run jarvis setupCheck the installation
Run diagnostics when a dependency, model or connection is missing.
uv run jarvis doctor04 / A CLEAR PICTUREGood to know
Good to know
before you begin.
Extension points are not preconnected services
MCP tools and skills describe what can be connected or taught. External accounts, dependencies and credentials still need setup.
Local-first is a data-path choice
Local files and settings stay on your machine until used by a configured tool or provider. Cloud models and connected services process the requests sent to them; voice requires Gemini Live.
The natural next questions.
Workflow images are generated illustrations, not screenshots of a connected assistant. Availability depends on your configured tools and providers.
Voice & two brains
Live conversation for the moment. A task agent for the work that follows.
Context & project memory
Recognize the project, remember its other name, and keep the next step close.
Browser & workspaces
Organize your real tabs, carry useful information across apps, and return to a saved setup.